机译的消歧策略  

A Strategy of Disambiguation in Machine Translation

作　　者：刘彬[1] 李丽[1] 

机构地区：[1]长沙理工大学外国语学院

出　　处：《中国科技翻译》2008年第1期24-27,F0003,共5页Chinese Science & Technology Translators Journal

基　　金：湖南省教育厅科研项目"对外交流过程中英汉语歧义结构比较及消歧策略"(编号:06C135)的部分成果

摘　　要：本文描述Transtar 3.0和DFKC2003英汉翻译软件在自动剖析源语中名词词组出现的歧义并进行归纳总结;为了保证译文通顺,提高翻译质量,尽量避免歧义的产生,文章通过客观分析比较,指出HPSG在名词词组的自然语言处理过程有更易于形式化的优势,建议采用基于HPSG理论的消歧策略,以改进目前机译名词词组不太令人满意的现状。This article describes and categorizes a number of ambiguous cases in the parsing processes of nominal phrases when using Transtar 3.0 and DFKC2003. As they adopt a context-free grammar or dependency grammar as the algorithms, the two pieces of software cannot operate smoothly and efficiently on the processes of parsing, marking, tagging, transmitting and outputting. It is recommended that in order to improve translation output, the software should execute the disambiguation strategies based on the theory of HPSG.

关 键 词：机器翻译 Transtar3.0 DFKC2003 自动剖析 HPSG 

分 类 号：H085[语言文字—语言学]